aermod version 13350 low wind options: sensitivity ... · may 2014 aermod version 13350 low wind...

May 2014

AERMOD Version 13350 Low Wind Options: Sensitivity Analysis and Evaluation Update

Study conducted on behalf of:

• API for sensitivity analysis• Lignite Energy Council for evaluation update

May 2014Environment Page 2

Outline of Presentation

• Purpose and objectives of this study for AERMOD low wind options

• Sensitivity Study Funded by API

– Model Options Tested

– Source Types Modeled for Flat and Complex Terrain

– Results of Sensitivity Analysis

• Evaluation Study Funded by Lignite Energy Council

– Models and Options Tested

– Database Tested to Date

– Results of Evaluation Tests to Date

May 2014Environment Page 3

Purpose and Objectives: Sensitivity Analysis

• Explore the sensitivity of the AERMOD low wind speed options for predicted impacts in both flat and complex terrain

• Tested for a variety of emission sources of interest to the American Petroleum Institute and their members

• 11 different source types were examined, ranging from tall buoyant point sources to low-level fugitive sources

• Examined types of sources significantly affected by use of the low wind options in AERMET and AERMOD

• NO2 was the pollutant selected assumed full conversion of NOx to NO2

• Model setup was based on hypothetical locations, but used input parameters and building downwash (when applicable) from real model applications

May 2014Environment Page 4

Model Configurations and Options for

Sensitivity Analysis

• AERMET/AERMOD Versions 13350 and 14134

• Three model configurations were run

1. AERMET/AERMOD all default

2. AERMET (Beta u*) / AERMOD (default)

3. AERMET (Beta u*) / AERMOD (LOWWIND2)

• Each model configuration was run for each source in both flat and complex terrain

• Results from AERMOD versions 13350 and 14134 were the same

May 2014Environment Page 5

Model Inputs: Source Types

Source ID Source Description

Stack

Height

(m)

Stack

Temp

(K)

Stack

Vel.

(m/s)

Stack

Diameter

(m)

FCCSource 1: a tall buoyant point source indicative of an FCC (fluid

catalytic cracking) refinery source (including building downwash)54.0 561.0 49.1 2.0

FLARESource 2: a tall buoyant point source representing a flare

(pseudo temp and velocity modeled to conserve buoyancy flux)75.6 1273.0 20.0 1.1

REGENHTR

Source 3: a tall buoyant point source indicative of a CCR

(continuous catalytic regenerative reformer) refinery source

(including building downwash)

104.2 450.0 12.2 3.7

GASTURBSource 4: a buoyant point source indicative of gas turbine at a

compressor station (including building downwash)13.7 777.0 41.6 1.2

DIESENGSource 5: a short-stack horizontal release point source indicative

of a diesel generator (including building downwash)9.1 697.0 0.001 0.60

DRILLRIGSource 6: a buoyant point source indicative of a drill rig (e.g.,

used at a fracking site, including building downwash)6.1 665.0 45.0 0.3

LNGTURBSource 7: a combustion turbine source indicative of drilling or

LNG facility operations.13.7 777.0 30.0 3.0

PNTTANKSource 8: a non-buoyant point source located on a tank (including

downwash)14.6 ambient 0.001 0.001

COMPRSTASource 11: buoyant point source associated with a compressor

station at a coal bed methane drilling site (including downwash)14.3 449.8 22.8 1.8

Release

Height

(m)

X-Dim.

(m)

Y-Dim.

(m)

Initial

Sigma-Z

(m)

AREA Source 9: a ground-level area source 0.0 50.0 50.0 0.0

Release

Height

(m)

Initial

Sigma-Y

(m)

Initial

Sigma-Z

(m)

ROADVOL Source 10: a volume source representing roadway traffic 10.0 14.0 16.0

May 2014Environment Page 6

Model Inputs: Receptors

Entire Grid

Near-field

• 50 meter spacing out to 500 meters

• 100 meter spacing out to 1,000 meters

• 200 meter spacing out to 2,000 meters

• 500 meter spacing out to 5,000 meters

• 1,000 meter spacing out to 10,000 meters

• 2,000 meter spacing out to 20,000 meters

• Additional receptors placed in

complex terrain areas

• AERMAP was used to calculate

elevation and critical hill heights for all

receptors

May 2014Environment Page 7

Model Inputs: Meteorology

• Two meteorological databases used in the study

1. Flat Terrain - 2007-2011 from Pascagoula, Mississippi

2. Complex Terrain - 2008-2012 from Page, Arizona

• Both meteorological databases feature a fairly large percentage of low wind speed hours

– Winds < 1.5 m/s at least 25% of the time

– Winds < 2.5 m/s at least 60% of the time

• The location of the hypothetical sources in complex terrain was strategically positioned near (and upwind of) a major terrain feature

May 2014Environment Page 8

Flat Terrain Wind Rose and Frequency Distribution

Wind speeds

< 2.5 m/s

over 60%

of the time

May 2014Environment Page 9

Complex Terrain Wind Rose

Wind speeds

< 2.5 m/s

over 60%

of the time

May 2014Environment Page 10

Model Relative Sensitivity Results: Flat Terrain

May 2014Environment Page 11

AERMOD Low Wind Sensitivity – Flat Terrain

• Tall buoyant stacks (FCC, FLARE, REGENHTR) were insensitive to the LW options - max impacts occur during unstable conditions

• Short buoyant stacks with downwash (DRILLRIG, COMPRSTA) insensitive to LW options - max impacts did not occur under light winds

• Short stacks without either momentum or buoyancy with downwash (DIESENG, PNTTANK) and fugitive sources are sensitive to LW options resulting in lower concentrations

– max impacts occurred under light wind stable conditions

– beta u* increase mechanical mixing and vertical dispersion

• LNGTURB (short buoyant non-downwashing) source experienced a high wind “side effect” of the LW options

– max impacts occur under high wind neutral conditions

– use of beta u* causes higher turbulence and plume touch down closer to the stack

• Low-level sources have peak impacts at low terrain near fenceline in complex terrain case as well

May 2014Environment Page 12

Model Relative Sensitivity Results: Complex Terrain

May 2014Environment Page 13

AERMOD Low Wind Sensitivity – Complex Terrain

• Tall buoyant stacks (FCC, FLARE, REGENHTR) are sensitive to the LW options in complex terrain

– For default options, max impacts occur under light wind speed stable conditions

– use of beta u* increases effective wind speed, mechanical mixing, vertical dispersion, and plume rise; reduces predicted concentrations

– use of LowWind2 also increases lateral dispersion and lower concentrations

• LNGTURB (short, non-downwashing) is sensitive to LW options

– For default options, max impacts occur under light wind stable conditions

– use of beta u* increases mechanical mixing height and vertical dispersion

• COMPRSTA (short, downwashing) responds to LW options

– For default options, max impact occurs under stable conditions (downwash)

– Lower max impact for beta u* option occurs in downwash during high wind unstable conditions

May 2014Environment Page 14 14

Lignite Energy Council Evaluation of

AERMOD Low Wind Options for Tall Stack Releases

North Dakota

Evaluation Study

Layout

May 2014Environment Page 15 15

Terrain

Contours

for SO2

Monitors

Used in

the ND

Study

(10-m

contour

interval)

May 2014Environment Page 16

Preliminary AERMOD Evaluation Results* with Actual

Hourly Emissions for North Dakota 4-year Database (07-10)

Monitor

Obs.

Conc.

4-yr Avg

99th %

Daily

Max,

µg/m3

AERMOD

14134

with

default

options:

Pre/obs

ratio

AERMET

with beta u*,

default

AERMOD:

Pre/obs ratio

AERMET with

beta u*,

AERMOD with

LOWWIND2

with min

sigma-v =

0.5 m/s:

Pre/obs ratio

DGC #12 91.52 1.28 1.28 1.05

DGC #14 95.00 1.45 1.45 1.05

DGC #16 79.58 2.00 2.00 1.58

DGC #17 83.76 2.07 1.49 1.29

Beulah 93.37 1.31 1.31 1.01

* Note: assumes SO2 background of 10 µg/m3

May 2014Environment Page 17

Conclusions

• This study reports sensitivity and field evaluation results for low wind options in AERMET/AERMOD

• Sensitivity was tested for 11 different source types

• In flat terrain, this option is important for low-level, non-buoyant source types, and not for tall, buoyant stacks

• In complex terrain, this option is very important for tall, buoyant stack releases

• Low wind speed evaluations are underway for real-world field databases featuring tall, buoyant stacks: ND, Gibson

• For the North Dakota database, low wind options lead to better AERMOD performance, especially for the elevated terrain monitor. SCICHEM does well for this database.